I agree actually. We built an open source agent-agnostic sandbox that we are landing improvements to, much more to come: https://github.com/anthropics/sandbox-runtime. That said the sandbox is not perfect, and any sandbox needs holes punched otherwise it is not useful in practice (at least for inference and MCPs, and almost certainly more). We will be doing much more in this area.

>and any sandbox needs holes punched otherwise it is not useful in practice

Don't assume our individual use cases. It can be perfectly useful with access to just one folder and zero network asside from the API.

That’s good to hear, it’s a much better approach than expecting humans to read commands and approve them or just trust that the LLM can set the proper horizons for its world itself while working.